16:55:48 #startmeeting 16:55:48 Meeting started Thu Mar 20 16:55:48 2014 UTC. The chair is regXboi. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:55:48 Useful Commands: #action #agreed #help #info #idea #link #topic. 16:57:39 #info dmm (paris) 16:57:55 #link https://wiki.opendaylight.org/view/TSC:Main#Meeting_Agenda 17:00:06 #info vijoy 17:00:21 #info Ryan Moats (regXboi) 17:00:25 #info Ed Warnicke :) 17:02:58 #info Chris Wright 17:03:01 (pdx) 17:03:46 are others in the WebEx? 17:03:56 and yes 17:04:10 colindixon: i'm dialled in 17:04:15 nothing being shared 17:04:29 vijoy: we aren't sharing any more 17:04:33 ah 17:04:45 ok, I'm in, just took a few minuts 17:04:53 #info Colin Dixon 17:04:53 vijoy: webex is a conf bridge now...irc for side discussion, meeting notes, and pasting links to things folks want to share 17:05:16 #info Chris Price joined the meeting 17:05:19 #info Madhu  17:05:53 #info Brent 17:06:08 +1 17:06:19 #topic remove minutes from meeting agent 17:06:35 #action acceptance of minutes will now be auto-on 17:07:20 #info RobDolin proxy for Rajeev Nagar (Microsoft) 17:07:20 #info skipping event updates (phil is not on the call) 17:07:35 #info skipping system integration/testing update (luis is not on the call) 17:07:51 #info skipping creation reviews(s) ... nothing on the list 17:10:25 #topic creation review(s) 17:10:38 #info discussion of the two week requirement for project reviews 17:11:11 eliminate the bureaucratic bureaucracy department of redundancy... 17:11:37 (heh, that's major delay from minutes == meetbot) 17:11:38 #action Madhu to schedule review of acceleration project to dmm 17:11:45 #topic committers 17:11:47 #agreed Madhu's proposal will be considered on Thu, Apr 3 17:12:17 #info consideration of Michal Polkorab as new committer on OpenFlowJava 17:12:54 #link https://lists.opendaylight.org/pipermail/openflowjava-dev/2014-March/000154.html - request to add Michal Polkrab as a committer to openflowjava 17:13:21 thanks RobDolin 17:13:24 #link https://lists.opendaylight.org/pipermail/openflowjava-dev/2014-March/000151.html - to the call for a vote on the openflowjava-dev mailer 17:13:55 #info Paul Zimmerman 17:15:59 #agreed Michal Polkorab will be added as a committer 17:16:06 #info David Meyer is asking the TSC to vote on approving Michal P. to committer 17:16:07 regXboi: Could you indicate to which project? 17:17:21 #info this is to the openflowjava project 17:17:40 #topic demo of the OpenDaylight Toolkit project 17:17:43 #info ... committer on OpenFlowJava project 17:18:58 #info Madhu is sharing screen for demo 17:19:21 #agreed that was committer to openflowjava 17:19:23 #link https://wiki.opendaylight.org/view/Project_Proposals:OpenDaylight_Toolkit the project proposal (which was e-mailed to TSC and project proposals lists) 17:19:35 #topic Madhu doing presentation on Archetypes 17:20:02 #link https://maven.apache.org/guides/introduction/introduction-to-archetypes.html maven documentation about archetypes 17:20:10 thanks colindixon :) 17:20:23 colindixon: heh, faster than my link pasting... 17:21:34 #info the general idea of the project is to provide skeletons for OpenDaylight applications which can be quickly spun up (and customized) by leveraging maven archetypes 17:21:57 #info it is informed by the rough edges found while Madhu, networkstatic and others were helping out running hackfests 17:31:02 #info this is very cool 17:31:35 #info archetypes 17:35:20 #info archetypes can be generated from existing code 17:36:17 #link http://maven.apache.org/archetype/maven-archetype-plugin/create-from-project-mojo.html 17:36:37 #info the archetype can also be extracted from an existing project 17:41:26 this is very neat! 17:41:44 #info colindixon stresses...eclipse built app w/ _no build errors_ (answers some large percentage of email questions via automation) 17:41:57 #info there is a no brainer 17:42:13 #info s/there/this 17:43:15 #info regXboi suggests this looks like a fundamental architectural shift (fundamental change to how we build ODL and its apps) 17:44:16 #info edwarnicke___ notes that they're really good for simple things, but they have more warts when you get to doing complicated things, so this is unlikely to extend to factoring out all modules in the current ODL codebase 17:44:35 regXboi this is not a shift on the architecture 17:44:47 rather this demo is just showing the power of modularity 17:44:55 we have already on ODL 17:44:55 GiovanniMeo: Agreed 17:45:05 so code generation has fundamental limitaions 17:45:13 but is not clear from building the controller as a monolith for example 17:45:43 WOW 17:46:26 So this is something that will sit outside of the project??? 17:46:38 As a way to leverage the project? 17:46:55 alagalah: it's a project proposal...so if we accept it, it'd be within the project 17:47:39 Okey dokey 17:47:54 my network performance is seriously degrading here. I'm going to use the IRC from here on out 17:47:58 alagalah: regardless, it's relevant to the project, and at worst a wonderful example of Forge space usage 17:48:06 (unless it gets better) 17:48:07 ok dmm, I'll get serious about scribing 17:48:10 #info even regXboi agrees that this is hugely useful for bringing new people into the project whether it's in hackfests or with new developers 17:48:22 colindixon - you beat me to it :) 17:48:24 regXboi: I can't #topic things as it turns out 17:48:27 cdub: All for that dude 17:48:31 ok, I'll topic then 17:49:18 cdub: its awesome stuff, just worried we are running faster than we are walking.... and making this consumable but I'll STFU 17:49:48 alagalah: this is like...we were running and incoming devs are still crawling 17:49:58 alagalah: here is tool to get them up and running 17:50:00 #info question from luis: can archetypes be used for apps on top of MD-SAL 17:50:15 cdub: lolz yeah agreed that I see the value, this is awesome 17:50:21 #info answer from Madhu: this is independent of MD-SAL/AD-SAL 17:50:31 #info there is already an archetype for model driven apps 17:50:33 #info so this is a good topic for MD-SAL 17:50:37 alagalah: Are you asking about where the archetypes are accessible? The 'archetype' gets stored in the nexus repo, and the stuff you generate them is just local (you could check it into git after you'e generated it :) ) 17:51:10 dmm: who do you want to call the next topic if you are IRC only? 17:51:23 yes, I will let you know 17:51:27 edwarnicke___: Well.... I thought cdub was saying this is a Forge thing 17:51:28 :) 17:51:55 alagalah: ah... I see :) 17:52:56 alagalah: no, just that it's a perfect use for ODLForge (esp. if we don't accept this as a project) 17:53:28 we can hear you just fine 17:53:46 cdub: No thats what I understood.... but still this is awesome, tools > examples 17:53:51 alagalah: in fact, it's an exact analog to PuppetForge (reusable puppet modules built by community for community that don't have to be part of core puppet release) 17:54:05 cdub: This is what I understood it as 17:54:09 ok, cool 17:54:13 #topic stable release (draft 1) 17:54:17 cdub: I can't imagine any reason we wouldn't accept this project ;) 17:54:29 edwarnicke___: ummmmm 17:54:45 Why not make a Forge for this ? 17:54:54 Or is that the project ? 17:55:03 alagalah: We are still trying to get the Forge infra up and stuff 17:55:38 alagalah: My point was in response to cdub's comment "no, just that it's a perfect use for ODLForge (esp. if we don't accept this as a project)" making the point that I can't see any reason we wouldn't accept it 17:56:37 edwarnicke___: Ok here's my point, Madhu has made something really cool, the lowers the barriers to entry, and I think this should be outside of the project, so folks don't have to jump on these calls to ask for permission to do stuff 17:56:57 #info plea from cdub for review of the high level process for stable releases 17:57:41 #info Ed Warnicke suggests we should get some wiki pages listing bugs that are good candidates for stable branch fixes 17:57:44 Hello folks, I turned on recording. I'm assuming that it wasn't off for a reason 17:58:03 phrobb: ugh... so we missed the archetypes demo 17:58:04 oh well 17:58:30 alagalah: that is not true 17:58:32 colindixon: no probb available 17:58:36 this is not a ODLForge project. 17:58:40 #info Ed Warnicke suggests we need a way to handle artifact versioning 17:59:01 dmm: I'm here now 17:59:06 #info stable release process...agree we want this, per-project stable branch, per-project committers to manage those branches, guidelines for what goes into that, releasing this 17:59:07 the archetype is a templating tool and other apps can be kickstarted from here 17:59:13 if that helps another project 17:59:26 they will move the app generated from the template to the ODLForge 17:59:31 ok, my mistake. welcome probb :-) 17:59:32 and work on the project creation 17:59:39 Ok, so I'll STFU 17:59:56 regXboi 17:59:57 #info Chris Wright suggests it will be helpful to have a document that helps committers say no to changes to a stable branch 18:00:01 i can help there 18:00:16 #info regXboi notes that we need to work on our ability to build snapshots of non-jar stuff so that we don't lose it as we release things. For example documentation and source code. 18:00:52 did ryan just say if we screw up opendove it doesn't matter :-) 18:01:20 dmm: no, he said that if we screw up opendove it doesn't break everyone else 18:01:26 dmm: we're a leaf project :p 18:01:27 dmm: compression is lossy...so yes you could argue he said that 18:01:49 Madhu: Ok clearly I got this wrong, I think this should be a project +1 18:02:21 But I like the idea of there being a "Forge" for ODL apps and this is an enabler right? 18:02:39 #info the git tag of the release 18:02:41 alagalah: everyone does and we're working on it (by "we" I really mean other people) 18:02:43 # is there 18:02:44 I think Forge is more of avoiding the red-tap 18:02:44 dmm: colindixon is right - if opendove gets messed up it's easy to fix and doesn't impact as many as if core gets messed up 18:03:04 while the tool-kit is to enable the developers to get into forge ASAP :) 18:03:10 Madhu: yes, and you know I'm not too bright and it takes me a while to put things together 18:03:28 alagalah: i beg to differ. u r in Noiro now :) 18:04:46 GiovanniMeo: thanks for the offer, I'll find you on #opendaylight to work on this 18:05:35 regXboi, yes you will find me there more often now 18:05:52 #topic stability, scalability, performance Hyrdogen releases 18:07:31 #info open question on candence 18:07:53 #link https://wiki.opendaylight.org/view/CrossProject:Stable_Release 18:09:01 just cut'n paste of email...formatting isn't perfect 18:11:03 #info release cadence discussion...3 dimensions: critical bugfix forces release, time based, queue length based 18:11:47 #info one request for time based to make it predictable for users (and developers -- know when the train is leaving the station) 18:21:49 #info discussion on what is a stable release seems to converge on "it's only stability/bugfix no new features" 18:22:57 #info RobDolin reporting 18:23:38 #info the criteria are guidelines for describing for what goes into the stable release, there are shades of grey here, judgement call needs to be made, we will make mistakes (chances are good at some point we'll introduce regression in stable release)...mitigation for this is transparency 18:23:55 #info a small group of folks are discussing documentation and are thinking of creating a documentation project for providing a clearing house of best practices for documentation as well as documentation tool chain 18:24:43 #info stable release patch proposal transparency can be done by literally posting the patch queue to the list (lkml style) or gerrit based mechanism for proposing patches to stable branches w/in a project 18:24:55 I know that mlemay has been working on things 18:24:58 http://docs.inocybe.com/install-guide/install-guide-20140316.pdf 18:25:01 http://docs.inocybe.com/install-guide/content/ch_install.html 18:25:03 #info we are gerrit based and more likely to adopt the latter model 18:26:42 colindixon: you want to link those? 18:27:02 #info mlemay (of inocybe) has been working on some documentation efforts 18:27:15 #link http://docs.inocybe.com/install-guide/content/ch_install.html html docs that are the current output 18:27:29 #link http://docs.inocybe.com/install-guide/install-guide-20140316.pdf pdf versions of the same 18:27:47 #link https://wiki.opendaylight.org/view/GettingStarted:Developer_Main - master list of IRC channels is here 18:27:49 #link https://wiki.opendaylight.org/view/CrossProject:Documentation_Group <- Cross-project documentation group as mentioned 18:28:30 #link https://wiki.opendaylight.org/view/IRC <- List of IRC channels 18:28:42 we also have: https://wiki.opendaylight.org/view/IRC 18:28:51 db coherency problem 18:29:32 #link https://wiki.opendaylight.org/view/Meetings 18:29:43 another db coherency problem 18:29:49 it's in the info box per project 18:29:51 RobDolin: Would you consolidate those lists? I don't really care which one you pick, as long as the one that goes away gets a pointer to the canonical one 18:31:10 #info colindixon suggests RSS feed w/ ical to put all meetings in one 18:31:58 creates a simple scheduling problem 18:32:21 the _only_ way to schedule is to go click on a bunch of links to figure out if your timeslot stomps on someone elses 18:32:37 * cdub prefers https://wiki.opendaylight.org/view/Meetings and https://wiki.opendaylight.org/view/IRC 18:33:00 cdub: +1 18:33:29 @cdub - Can those .../view/... pages be auto generated ? 18:34:21 RobDolin: have to ask tykeal , i don't know how to autogenerate with mediawiki 18:35:17 this is what i'm used to: https://wiki.openstack.org/wiki/Meetings 18:35:34 and it does include ical feed 18:35:49 #info colindixon suggests that we write code for a living 18:36:13 #action colindixon to look into how to coordinate meetings 18:37:00 #info colindixon to look into this 18:37:43 #topic open election status by phrobb 18:37:45 regXboi: thanks, I'll pull in edwarnicke___ and cdub for a bit of help, but the idea that we're going to make things harder because we can't reliable copy information between two wiki pages makes me hurt on the inside 18:39:23 regXboi: it's FUD (not in the smear campaign sense...just lack of comfort) 18:39:50 regXboi: and i don't agree w/ the decision...but hey...consensus is as consensus does 18:39:53 https://meetings.opendaylight.org/opendaylight-meeting/2014/opendaylight-meeting.2014-03-13-16.59.html bullet #15 18:39:53 * regXboi_ lost my link again 18:40:04 regXboi_: it's FUD (not in the smear campaign sense...just lack of comfort) 18:40:09 regXboi_: and i don't agree w/ the decision...but hey...consensus is as consensus does 18:40:25 #info regXboi asked why only committers? 18:40:42 #info answer was to keep things sane 18:41:43 #info question is how to remove oneself from the nomination list 18:42:13 #action phrobb will send email to nominees to get confirmation that they want to run 18:42:26 #info question on what is the term being elected for 18:42:33 #info answer is one year 18:43:51 #topic call for new topics 18:44:02 #info question if there will be a meeting next week 18:44:15 #action dmm will propose to cancel next weeks meeting 18:44:53 #endmeeting